This article is the second in a four-part series exploring the transformative role of the Digital Data Steward. In the first article, we introduced the concept of the Digital Data Steward, laying the foundation for how a system of agents within a coordinated framework can revolutionize the role of a Data Steward.

This article builds on that introduction by focusing on the Data Strategy Agents. These specialized AI agents support defining and executing a domain-specific data strategy.

---

What are Data Strategy Agents?

Data Strategy Agents are specialized, AI-augmented agents that assist in defining a data strategy, and then perform tasks and actions that operationalize the data strategy. As these agents typically operate at the domain level, we call them Domain Data Strategy Agents (DDSA).

This domain strategy is part of the overall enterprise data strategy defined by the Chief Data Officer (CDO). The domain strategy would include: 

1. Identifying domain-specific critical data elements (master data, transactional, reference).  
2. Defining data strategy themes and new high-level domain data capabilities necessary to enable business outcomes and objectives. Data capabilities include technical, data process, quality, regulatory, and governance.
3. Prioritizing/classifying/quantifying those domain critical data elements and data capabilities based on a strategic assessment impact.  
4. Defining the Key Performance Indicators (KPIs) and business metrics to track the domain data strategy success. 

The Domain Data Strategy Agents are among the most complex and nuanced agents in the Digital Data Steward agentic system. It is best to think of them as important members of the overall data strategy team. The Domain Data Strategy Agents can handle the heavy lifting — analyzing inputs, proposing capabilities, creating strategic data themes, simulating scenarios, tracking KPIs, and generating dashboards. However, it will still be necessary to validate the outcomes and provide implementation details.

Human oversight is essential at key judgment points, especially where nuance, risk tradeoffs, or organizational politics and complexity are involved. Providing the initial deliverables, backed by industry and internal facts, will greatly accelerate the development of the final strategy and its adoption by the business teams.

How do Domain Data Strategy Agents help build data strategy? 

Figure 1: Building the data strategy with AI agents

1. Align data strategy with business objectives

The agents help develop a domain data strategy based on input from various internal and external sources and apply intelligence to listen with “data ears” for the implied and explicit strategic data themes and high-level data capabilities required.

The agent will use a suite of AI-driven capabilities and other tools, such as LLMs, to identify strategic data themes, domain-specific critical data elements, and data capabilities. Then, it will construct an attribution map that links strategy insights to their sources — enhancing transparency, proof points, and understanding of trend trajectories.

Many of these tools exist today and are ready for use. Others will require development, either by the organization, solution providers, or startups, offering collaboration potential for CDOs.

Training a Domain Data Strategy Agent to “listen with data ears” like an experienced strategy team of domain human experts must be grounded in a rich corpus of strategic and operational knowledge. The information needed to train the models will come from a variety of both internal and external sources.

This includes curated training sets built from CDO/Data Steward playbooks, industry frameworks (such as DAMA-DMBOK and DCAM), regulatory guidelines, real-world data strategy case studies, and best practices from leading organizations. The table below suggests a number of sources that could be ingested by the Domain Data Strategy Agent to formulate the data strategy requirements and objectives.

These materials enable the agents to recognize patterns in executive language, infer the underlying business intent, and recommend relevant data capabilities. By learning from how seasoned data experts interpret and act on strategic goals, the agent develops the contextual reasoning needed to translate ambiguous business direction into precise, prioritized data initiatives.

Table 1: Internal and external data sources the Domain Data Strategy Agent can ingest to formulate data strategy requirements and objectives    

2. Analyze current state of data capabilities 

Using the output from the first phase, the Domain Data Strategy Agents will help with assessing the current data capabilities, specifically:

• Inventorying existing critical data assets in their domain (data repositories)

• Evaluating data quality metrics (accuracy, completeness, consistency)

• Evaluating data flows, data error logs and process bottlenecks

• Rationalizing the classification models used by different data “voices” for their domain-critical data elements (security, privacy, CDO classifications)

3. Define data requirements for high-level capabilities

Using published data capability maps, knowledge graphs and ontologies linking specific datasets and elements to business outcomes, capabilities and goals, the Domain Data Strategy Agents can assist with outlining high-level capabilities required for:

• Data quality, governance, and security

• Privacy and regulatory standards

• Technical capabilities

4. Assess strategic impact

Using sentiment analysis, industry trends, historical models and predictive learning, the Agent will assign a strategic impact weighted assessment score to each of the data capabilities and critical data elements.

The scoring model includes various dimensions and a scoring percentage adjusted with input from the business and later with reinforcement learning:

• Customer impact 

• Process and productivity impact
• Risk impact

• Financial impact and ROI

• Implementation dependencies and complexity

5. Monitor and refine data strategy dynamically  

The output from these tasks is a prioritized list of Common Data Elements and high-level data capabilities for each of the strategic themes identified. A visual dashboard showing where these data requirements fall across the complexity axes via a strategic impact heatmap is provided.

These models may be continually refined based on changes in business priority, customer feedback, regulatory changes, and KPI progress. The progress is tracked in collaboration with the other agents in the Digital Data Steward agentic system based on data strategy KPIs, and metrics based on strategic business value and capability execution.

The Domain Data Strategy Agent predicts reasonable target values for KPIs based on industry benchmarks and internal baselines. Adjustments can be made based on company size, complexity, and regulatory environment. Later, user feedback and real-world progress can be used to further refine KPIs dynamically (via reinforcement learning models).

The Domain Data Strategy Agent proposes four main categories of KPIs: 

1. Execution KPIs to track specific activities tied to strategy implementation. Example: % Data Capabilities enabled, % Critical Data Elements documented. 
2. Outcome KPIs to measure actual business or operational improvements.
   Example: Reduction in customer churn linked to better CRM data quality.
3. Data Health KPIs to assess the intrinsic health of the data ecosystem.
   Example: Data Quality Score, Metadata Completeness, % Automated Data Flows.
4. Maturity Progress KPIs to map progress across maturity stages.
   Example: % Maturity Milestones Achieved, % Governance Frameworks Adopted. 

How do Domain Data Strategy Agents work with other AI agents in the Digital Data Steward system? 

The Domain Data Strategy Agents (DDSA) collaborate with other agents in the Digital Data Steward agentic system. The DDSAs provide strategic themes and input to these other agents to develop their rules and processes. Additionally, they continually monitor external reporting, industry trends, and internal factors to adjust the data strategy as necessary and take action to alert these strategy changes to other agent tasks and processes.

Together, these interactions help maintain a seamless, cohesive execution of the data strategy while ensuring all agents are aligned with strategic goals.  

To that end, we also see the need for an “Agent Shepherd,” in an orchestration role to be added to the Digital Data Steward agentic system. This role can be taken over by a separate agent, or these responsibilities and/or tasks can be given to the overall Digital Data Steward. Later, as the organization matures, a separate enterprise-wide Agent Shepherd at the CDO level may be deployed.

Table 2: How an AI Agent Would Execute to Support Data Strategy Building

Conclusion 

Leveraging Domain Data Strategy Agents represents a foundational shift in the future of data strategy development. By blending AI with domain expertise and contextual awareness, the Domain Data Strategy Agents accelerate the domain data strategy development and organizational buy-in. They industrialize CDO intuition by “listening,” reasoning, and constantly reprioritizing to keep data investments locked on current trends and business value.

As a crucial part of the Digital Data Steward agentic system, Domain Data Strategy Agents work in collaboration with the other agents — specialized in data quality, master data management, or compliance, but equally important with the CDO, business, and IT experts — to complete and validate the domain strategy.

In the next article, we will explore how AI agents transform data quality, metadata, and master data management — another vital aspect in the Digital Data Steward ecosystem.

About the Authors:

Maria C. Villar brings over 30 years of experience as a transformational technology executive, having served as Chief Data Officer in both the technology and financial sectors. Currently, she is Co-founder and Managing Partner of Business Data Leadership, a firm committed to enhancing effective data and AI management practices through training, writing, coaching, and consulting. Her expertise includes enterprise data strategy, data and AI governance, business value realization, organization and change management, and ESG and Sustainability.

Recognized as a leader in the data and AI industry, Villar is a frequent speaker and author. Her accomplishments include co-authoring the book “Managing Your Business Data from Chaos to Confidence” with Theresa Kushner, developing online master classes, e-learning modules, and webinars, contributing to “Latin Business Today” since 2010, and serving as the WLDA Ventures Program Manager for an accelerator program focused on data and AI startups.

Mike Alvarez is a data and AI transformation leader with over 20 years of experience driving innovation at the intersection of data science and commercial product development. He helps organizations unlock transformative value from their data, technology, and human resources. His career spans pioneering data leadership roles at Fortune 20 companies where he delivered hundreds of millions in business value through data/AI initiatives.

As CTO and Head of Product at NeuZeit, he is focused on accelerating the value and adoption of AI for organizations with acceleration frameworks. Alvarez is passionate about helping companies navigate their data and AI transformation journey by establishing robust data foundations, deploying scalable AI solutions, and creating platforms that democratize insights to drive competitive advantage. Mike is also a board member of the AI Freedom Alliance (https://aifalliance.org/) advocating for the fair and ethical use of Artificial Intelligence.

Elizabeth (Beth) Hiatt is Head of Global Data Governance at PayPal. She has close to 30 years of experience building and deploying enterprise-wide data management and governance programs. Beth has held various data management and governance roles across business and technology in financial services, telecommunications, and hospitality. She has implemented enterprise data management programs end-to-end, developing and enabling critical functions such as data governance, data quality, and master and metadata programs. She has deep technical expertise in enterprise data architecture, helping organizations “connect the dots” across the data lifecycle.

Beth is a strong, results-driven leader with experience managing large, complex organizations specifically focusing on growing a company’s data management maturity while changing the organization’s data culture. She has written articles including “Time to Level Up: The Evolving Role of the Chief Data Officer” published by TDWI, spoken at many conferences including the Women Data Leaders Global Summit in 2021, and was on CDO Magazine’s Global Data Power Women List in 2022.

Christine Legner is a Professor of Information Systems at the Faculty of Business and Economics (HEC), University of Lausanne, in Switzerland. Her research fields are data management, enterprise architecture, and business software. She is the co-founder and academic director of the Competence Center Corporate Data Quality (CC CDQ), an industry-funded research consortium and expert community dedicated to advancing the field of data management. In this role, Legner leads a research team that collaborates closely with industry experts from 20 Fortune 500 companies (BASF, Bayer, Bosch, Nestlé, Schaeffler, SAP, Siemens, and Tetrapak, among others) to develop innovative concepts, tools and methods for data management.

Together with Dr. Richard Wang, Legner also serves as the Co-Chair of the annual CDOIQ European Symposium, which brings together CDOs, CAOs, CAIOs, and senior leaders shaping the data, analytics, and AI landscape in Europe.